What I suggested doesn’t contradict what you’re saying.
I just suggested to divide our waiting for Roon MQA to 2 phases:
It is speculation on my part but which bit do think is incorrect, the whole statement or just a part?
Fair enough… My understanding is that MQA is “fully decoded” prior to any DSP. Currently the only implementation that has a meaningful DSP post full MQA decoding is the Meridian DSP speakers, where room correction DSP is applied post full decoding on the final PCM data stream. This makes perfect sense. Paraphrasing the MQA literature, MQA full decoding gets you the most accurate PCM representation of the original sound, including your DAC’s characteristics. Once full decoding is achieved to the max rate of the DAC in question, DSP is applied.
Yes I agree we are saying the same thing. I just added that my understanding is that Roon would like to go past the first unfold and do a second “generic” unfold given the MQA embedded info post the first unfold specifies these filter settings for each track.
Consider for example the Dragonfly: The input controller firmware was updated to understand this embedded MQA data and interpret the MQA-specific settings into ESS-specific settings for the ESS DAC. The DAC firmware was untouched.
In Roon one could interpret the MQA-specific data embedded in the PCM stream coming out of the first unfold to set upsampling parameters in Roon’s upsampling algo. Once upsampled in Roon to the max rate allowed by the DAC, you can do further DSP in Roon to apply eq, room correction, etc.
The argument against this is that a DAC-specific implementation is better, but I don’t think this is an argument that can stand any scrutiny. Consider the key factors in the sound you get: analog stages, speakers, and room. All of these are clearly stronger drivers of sound quality than the differences we are talking about, and having the ability to apply eq or room corrections is much more important in these cases.
(FWIW, I don’t use room corrections myself, but a lot of people do)
Nope, not even Meridian has an MQA + room correction solution. All that SE DSP loudspeakers have is treble tilt, bass, and balance.
I agree with what you said.
It’s just that whatever holy grail of a solution Roon is trying to push for in the future, I’m sure that the first unfold is already developed (and the Roon devs probably use it in their private builds 
) and could be released tomorrow. And I don’t see why MQA would have a problem with that, since Tidal and A+ are already doing it. Maybe Roon are worried about people expecting more and complaining about it. But I personally think that is neglible and can be solved with being clear about it in the release notes and the signal chain.
I can only share this … hearing John Coltrane’s (almost) entire catalog in MQA via Tidal/Audirvana+ has been a significant improvement in my enjoyment of that music. Same goes for the now numerous 60’s Blue Note albums (Herbie Hancock, etc) available in MQA …and more. Btw, I say this as someone who owns MANY of Coltrane’s albums, 60’s Blue Note albums, etc in 24/196 or 24/96 from HDtracks. With Audirvana+ now providing ‘network’ capabilities, I can send MQA decoded files to my ultraRendu => Kef LS50w speakers. Result? … my use of Roon has dropped by 75%. Audirvana also supports Qoboz, so I’m asking myself…do I care about my user experience (Roon) more than I do sound quality?
If Roon and MQA do not see a partnership in their future, ** I ** would greatly appreciate Roon sharing that info, so I can sell my Sonic Transporter / Roon server and move on. Reading in this thread …
“MQA is literally the last topic on earth that would provoke us to break that longstanding practice. There’s way too much attention and anticipation around it” …
is extremely disappointing, at least, for me. Luckily for me, with new streaming services starting to offer MQA (Deezer, etc), with more to come…HDmusicStream, Quboz, etc…there are good options outside of Roon/Tidal. Maybe I just don’t ‘get it’, but it seems to me, the fast growing embrace of MQA makes it in Roon’s best interest to move MQA implementation waaaay higher on their priority list versus thinking there’s “way too much attention and anticipation around it”. YMMV.
The context there is that the Roon guys aren’t going to predict a release date for an MQA compatible version, not that they aren’t pursuing MQA compatibility, and not that it isn’t a priority.
Count me among the users who’d like to see it happen sooner rather than later, but I’ve come to appreciate the Roon team’s notion that whatever features are added should be best in class. So I’ll just wait for that, albeit somewhat impatiently.
I think what Brian was saying there is that there are so many on eyes on this feature that, of all the features, if they gave a release date it would cause the most mahem if things slipped a little, due to having more attention than anything. He’s not saying it’s low priority at all. I think it’s actually high priority but they’re trying to do it right.
Brian hinted in another thread that getting the ‘M’ symbol to show when browsing Tidal (like the Tidal desktop app) is quite complicated. I wouldn’t be surprised (only guessing) if Core decoding and the ‘M’ symbol both debut at the same time and that’s why it’s taking a little longer.
You like the MQA better than the 24/192 from HDtracks?
yep. Sounds more ‘analog’ to my ear. Plus the attack on acoustic instruments is more accurate, especially if you like an ‘upfront’ (musician’s perspective) presentation. As a former studio musician, etc…it sounds strikingly more accurate.
Ahhh…I was mis-reading his comment…thank you!
Is this using the same transport and the same DAC?
No problem, I can see how it could be interpreted the other way.
But the most viewed thread on this entire forum is the “MQA General Discussion” thread, so that’s how I guessed what Brian meant, regarding attention and anticipation.
My comparisons have used Audirvana+ 3.1.2 delivering hires files from my NAS and decoding (first unfolding) MQA from Tidal to my sonore ultraRendu => DAC inside my LS50w speakers. Just to be clear, there may be a few aspects of the hires files I find preferable, but overall, MQA is my preference.
What has been your experience with MQA?
I agree with you as to the ideal implementation.
I suspect one problem is those dratted lights on various MQA compatible DACS. Because MQA holds itself out as an “authentication” process I suspect those lights won’t light up if post MQA final DSP is applied prior to the DAC. MQA may have a problem with that.
The only way I can see is if Roon re-inserts the appropriate MQA flag in the stream after ‘first unfold’ and DSP treatment. An MQA DAC would then treat the stream as ‘authenticated’ and do further processing if required. This may be impossible due to MQA preventing it of course. I’d hate to be left out of the MQA game because I use DSP on the stream.
Those embedded meta data as well as flags don’t seem to withstand any form DSP or bit perfect conversion. Simple DSP such as gain settings will disrupt the authenation and bit perfect conversion such as converting from MQA FLAC to WAV or back to MQA FLAC loses its ability to decode.
Bit-perfect conversion does NOT affect correctly implemented MQA decoding.
Ha! Very good point! I had not factored that in.